TP1_Reinforcment_Learning_Py/gridrender.py at master · AntoinePlumerault/TP1_Reinforcment_Learning_Py · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
from tkinter import *
import numpy as np


class GUI(Canvas):
    def __init__(self, master, *args, **kwargs):
        Canvas.__init__(self, master=master, *args, **kwargs)


def draw_square_q(polygon, x, y, q, actions, dim=50):
    polygon.create_polygon([x, y, x + dim, y, x + dim, y + dim, x, y + dim], outline='black',
                           fill='white', width=2)

    font = ('Helvetica', '30', 'bold')

    for i, a in enumerate(actions):
        if a == 0:
            polygon.create_polygon([x + dim, y, x + dim / 2., y + dim / 2., x + dim, y + dim], outline='gray',
                                   fill='red', width=2)
            polygon.create_text(x + 3 * dim / 4., y + dim / 2., font=font, text="{:.3f}".format(q[i]), anchor='center')
        elif a == 1:
            polygon.create_polygon([x, y + dim, x + dim / 2., y + dim / 2., x + dim, y + dim], outline='gray',
                                   fill='green', width=2)
            polygon.create_text(x + dim / 2., y + 3 * dim / 4., font=font, text="{:.3f}".format(q[i]), anchor='n')
        elif a == 2:
            polygon.create_polygon([x, y, x + dim / 2., y + dim / 2., x, y + dim], outline='gray',
                                   fill='yellow', width=2)
            polygon.create_text(x + dim / 4., y + dim / 2., font=font, text="{:.3f}".format(q[i]), anchor='center')
        elif a == 3:
            polygon.create_polygon([x + dim, y, x + dim / 2., y + dim / 2., x, y], outline='gray',
                                   fill='purple', width=2)
            polygon.create_text(x + dim / 2., y + dim / 4., font=font, text="{:.3f}".format(q[i]), anchor='s')


def draw_square_policy(w, x, y, pol, actions, dim=50):
    w.create_polygon([x, y, x + dim, y, x + dim, y + dim, x, y + dim], outline='black',
                     fill='white', width=2)

    font = ('Helvetica', '30', 'bold')
    if (hasattr(pol, "size") and pol.size > 1) or isinstance(pol, list):
        d = pol
    else:
        d = [-1] * len(actions)
        idx = actions.index(pol)
        d[idx] = 1

    for j, v in enumerate(d):
        if j < len(actions):
            a = actions[j]
            if a == 0 and v > 0:
                w.create_line(x + dim / 2., y + dim / 2., x + 3*dim / 4., y + dim / 2., tags=("line",), arrow="last")
                if not np.isclose(v, 1.):
                    w.create_text(x + 3*dim / 4., y + dim / 2., font=font, text="{:.1f}".format(v), anchor='w')
            elif a == 1 and v > 0:
                w.create_line(x + dim / 2., y + dim / 2., x + dim / 2., y + 3* dim / 4., tags=("line",), arrow="last")
                if not np.isclose(v, 1.):
                    w.create_text(x + dim / 2., y + 3*dim / 4., font=font, text="{:.1f}".format(v), anchor='n')
            elif a == 2 and v >0:
                w.create_line(x + dim / 2., y + dim / 2., x+dim/4., y + dim/2., tags=("line",), arrow="last")
                if not np.isclose(v, 1.):
                    w.create_text(x + dim / 4., y + dim / 2., font=font, text="{:.1f}".format(v), anchor='e')
            elif a == 3 and v >0:
                w.create_line(x + dim / 2., y + dim / 2., x + dim / 2., y + dim / 4., tags=("line",), arrow="last")
                if not np.isclose(v, 1.):
                    w.create_text(x + dim / 2., y + dim / 4., font=font, text="{:.1f}".format(v), anchor='s')


def render_q(env, q):
    root = Tk()
    w = GUI(root)
    rows, cols = len(env.grid), max(map(len, env.grid))
    dim = 200
    w.config(width=cols * (dim + 12), height=rows * (dim + 12))
    for s in range(env.n_states):
        r, c = env.state2coord[s]
        draw_square_q(w, 10 + c * (dim + 4), 10 + r * (dim + 4), dim=dim, q=q[s],
                      actions=env.state_actions[s])
        w.pack()
    w.pack()
    root.mainloop()


def render_policy(env, d):
    root = Tk()
    w = GUI(root)
    rows, cols = len(env.grid), max(map(len, env.grid))
    dim = 200
    w.config(width=cols * (dim + 12), height=rows * (dim + 12))
    for s in range(env.n_states):
        r, c = env.state2coord[s]
        draw_square_policy(w, 10 + c * (dim + 4), 10 + r * (dim + 4), dim=dim, pol=d[s],
                           actions=env.state_actions[s])
        w.pack()
    w.pack()
    root.mainloop()